Short Stories

A novel peptide CXCR ligand derived from extracellular matrix degradation during airway inflammation

A novel peptide CXCR ligand derived from extracellular matrix degradation during airway inflammation
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  A novel peptide CXCR ligand derived from extracellularmatrix degradation during airway inflammation Nathaniel M Weathington 1 , Anneke H van Houwelingen 2 , Brett D Noerager 1 , Patricia L Jackson 1 ,Aletta D Kraneveld 2 , F Shawn Galin 1 Gert Folkerts 2 , Frans P Nijkamp 2 & J Edwin Blalock  1 We describe the tripeptide neutrophil chemoattractant N-acetyl Pro-Gly-Pro (PGP), derived from the breakdown of extracellularmatrix (ECM), which shares sequence and structural homology with an important domain on alpha chemokines. PGP causedchemotaxis and production of superoxide through CXC receptors, and administration of peptide caused recruitment ofneutrophils (PMNs) into lungs of control, but not CXCR2-deficient mice. PGP was generated in mouse lung after exposure tolipopolysaccharide, and  in vivo   and  in vitro   blockade of PGP with monoclonal antibody suppressed PMN responses as much aschemokine-specific monoclonal antibody. Extended PGP treatment caused alveolar enlargement and right ventricular hypertrophyin mice. PGP was detectable in substantial concentrations in a majority of bronchoalveolar lavage samples from individuals withchronic obstructive pulmonary disease, but not control individuals. Thus, PGP’s activity links degradation of ECM with neutrophilrecruitment in airway inflammation, and PGP may be a biomarker and therapeutic target for neutrophilic inflammatory diseases. Neutrophils are effectors in pulmonary inflammatory diseases, includ-ing chronic obstructive pulmonary disease (COPD) and acute lunginjury  1 . PMNs are required for pulmonary symptoms arising fromtrauma and hemorrhage or lipopolysaccharide (LPS), as their deple-tion eliminates the phenotype 2 . The major chemoattractants forneutrophils during inflammation are Glu-Leu-Arg motif–containingELR  + CXC chemokines 3 including IL-8 (CXCL8) and GRO- a , GRO- b and GRO- g  (CXCL1, CXCL2 and CXCL3, respectively) in humansand KC and MIP-2 (CXCL1 and CXCL2) in mice. The receptors forthese molecules are CXCR1 and CXCR2 in humans 4,5 and only CXCR2 in mice 6 .Others have reported chemotactic activities of ECM- or collagen-derived peptides 7–10 without description of a molecular mechanismfor such activity. The PGP peptide was discovered in a rabbit modelinvestigating alkali injury to the eye, where PMN inflammationleads to corneal ulceration and perforation. It was shown thatalkali treatment of cornea generates PGP, probably from collagen 11,12 ;the peptide is chemotactic for neutrophils; and corneal peptidetreatment causes neutrophilia similar to that seen in alkali injury  13 .Similarly, collagen peptides are active in lung. Intratracheal admin-istration of collagen fragments causes accumulation of pulmonary neutrophils 14 , whereas overexpression of collagenase-1 (ref. 15) orblockade of matrix metalloproteinases (MMPs) 16 causes emphysemaand attenuates symptoms of COPD, respectively. Airway exposure of mice to LPS causes a pathology in which MMP-9 activity and lungneutrophil burden correlate highly  17 .Here we sought the specific molecular mechanism of PGP’s activity on neutrophils and evaluated its relevance to lung pathology. Westructurally compared PGP with CXCL, characterized PGP’s activity on human and mouse PMNs and CXCR transfectants, and testedreceptor binding.  In vivo , we tested PGP’s effects on airway celldistributions in normal and CXCR2-deficient ( Cxcr2 –/– ) mice, andcharacterized the presence and activity of PGP  in situ  in pulmonary inflammation caused by exposure to LPS. We tested whetherlong-term exposure to PGP caused pathology by monitoringalveolar enlargement and right ventricular hypertrophy. Finally, weassayed BAL samples from individuals with COPD and controlindividuals for PGP. RESULTSPGP is chemotactic for neutrophils  in vivo   and  in vitro  In initial  in vivo  experiments, we exposed 6–8-week-old C57 Bl/6Jmice to synthetic, highly purified, endotoxin-free PGP or controlpeptide Pro-Gly-Gly (PGG) in sterile hospital-grade phosphate-buf-fered saline (PBS) or PBS alone by intratracheal administration. Wekilled the mice 5 h later, performed bronchoalveolar lavage (BAL), anddifferentially counted airway cells. PGP caused a marked increase inthe total number of cells in lungs of PGP- but not PGG-exposed mice( Fig. 1a ); this increase is completely accounted for by the rise inneutrophil numbers, with other cell types unaffected by either peptide,and BALB/c mice responded similarly (data not shown). Thus, PGPcauses specific PMN recruitment into the airways.To systematically evaluate the effects of PGP on neutrophils,we used a transwell chemotaxis system with purified PMNs orneutrophil-like differentiated HL-60 cells placed in media above a3  m m polycarbonate filter, and PGP, chemokines or media placed in Received 19 September 2005; accepted 27 December 2005; published online 12 February 2006; doi:10.1038/nm1361 1 Department of Physiology and Biophysics, University of Alabama at Birmingham, 1918 University Boulevard, Birmingham, Alabama 35294, USA.  2 Utrecht Institutefor Pharmaceutical Sciences, Department of Phamacology and Pathophysiology, Utrecht University, Sorbonnelaan 16, 3584 CA Utrecht, The Netherlands.Correspondence should be addressed to J.E.B. ( NATURE MEDICINE  VOLUME 12  [  NUMBER 3  [  MARCH 2006  317 ARTICLES    ©   2   0   0   6   N  a   t  u  r  e   P  u   b   l   i  s   h   i  n  g   G  r  o  u  p    h   t   t  p  :   /   /  w  w  w .  n  a   t  u  r  e .  c  o  m   /  n  a   t  u  r  e  m  e   d   i  c   i  n  e  the lower chamber. After 1 h, we quantified migration and indexed itto basal (media control) activity. PGP was chemotactic for neutrophilsbetween 10 nM and 100  m M, and PGG had no activity, consistent withthe  in vivo  data ( Fig. 1b ). PGP was active on purified PMNs fromhumans and mice as well as HL-60 cells 18 , with 50% effectiveconcentration (EC 50 ) values from 0.5 to 6  m mol for all cells tested( Fig. 1c ). PGP is chemotactic rather than chemokinetic, as disruptionof the PGP gradient between chambers suppressed cell migration.IL-8–dependent chemotaxis was similarly disrupted by PGP in theupper chamber ( Supplementary Fig. 1  online). PGP has structural homology to CXC chemokines Searching for a mechanism for the activity of PGP and other collagen-derived peptides, we looked for similarity between them and theCXC chemokines classically defined as inflammatory neutrophilchemoattractants. Sequence comparison of a collagen fragmentisolated using the 9A4 antibody, which recognizes the PPGPQsequence 19 , to chemokine sequences showed that several ELR  + CXC chemokines (CXCL1, CXCL2 and CXCL3 of human, mouseand rat) contain a conserved PPGPH sequence immediately N-terminal to the third structural cysteine. IL-8 has the sequenceESGPH in this position ( Fig. 2a ), and none of the ELR  – CXC or CCchemokines possess this ‘GP’ motif seen in all neutrophil-specificchemokines. Earlier structure-function studies of IL-8 showed thatthis GP motif was an absolute requirement for neutrophil cell bindingand activation in radioligand and elastase release assays, respectively  20 .We next compared the available structures of IL-8 (ref. 21) andPGP 22 and found that the molecular orientation of the SGP motif of IL-8 is well represented in the structure of the predominant ‘all- trans ’isomer of PGP ( Fig. 2b , c ). In IL-8, the SGP motif ( Fig. 2d ) follows thefirst of three  b -strands and resides in the so-called ‘30s loop’ of themolecule, which is in close proximity to the ELR domain ( Fig. 2d ).These domains are exposed to solvent and linked by the bridgingdisulfide immediately C-terminal to each ( Fig. 2d ). Moreover, theSGP region of IL-8 is alone active as a chemoattractant with dosimetry similar to PGP ( Supplementary Fig. 1  online). PGP activity is mediated by CXCR1 and CXCR2 To test whether PGP and IL-8 act on a common receptor, we usedmonoclonal antibodies targeting the CXCR1 and CXCR2 receptorsabundant on human PMNs. Incubating PMNs with a mixtureof antibodies to these receptors predictably suppressed IL-8–dependent chemotaxis to control levels ( Fig. 3a ). When we evaluatedPGP-dependent chemotaxis, pretreatment with the same antibodies Figure 1  PGP is chemotactic for PMNs  in vitro  and  in vivo  . ( a ) The PGP peptide, but not thePGG control peptide causes an increase in airwayPMNs after intratracheal administration (250  m gin 50  m l) to 6–8-week-old C57 BL/6J mice; BALand differential cell counts were performed 5 hafter treatment (* P  o 0.05 compared to PBSgroup). ( b ) PGP is a chemoattractant of humanPMNs in an  in vitro   transwell chemotaxis assay,whereas PGG had no chemotactic activity; 2   10 5 PMNs were placed on one side of a 3  m mfilter with peptides present on the other side, and migrated cells were photographed and quantified after 1 h at 37  1 C. ( c ) Dose-response curves for humanPMNs (open squares), HL-60 cells (filled squares), bone marrow–derived mouse PMNs (open triangles) and mouse peritoneal PMNs (filled triangles).   0 .  0  0  1 PGP ( µ M)Peptide ( µ M)   0 .  0  1  0 .  1  1  0  1  0  0  1 ,   0  0  0 020406080100    C   h  e  m  o   t  a  c   t   i  c   i  n   d  e  x   P  e  r  c  e  n   t  m  a  x   i  m  u  m  c   h  e  m  o   t  a  c   t   i  c  r  e  s  p  o  n  s  e   C  e   l   l  s   /  m   l    ×     1   0    4   1 PBSPGPPGPPGGPGG   0  0 .  0  0  1 1234   0 .  0  1  0 .  1 1  1  0 Total cellsPMNs a b c ** 806040200 COLLAGEN FRAGMENT  DDGPSGAEGPPGPQGLAGQR 1. Rat CINC (CXCL1)  MVSATRSLLCAALPVLATSRQATGAPVANELRCQCLQTVAGIHFKNI*QSLKVMPPGPHCTQTEVIATLKNGREA*CLDPEAPMVQKIVQKMLKGVPK  2. Mouse KC (CXCL1)  MIPATRSLLCAALLLLATSRLATGAPIANELRCQCLQTMAGIHLKNI*QSLKVLPSGPHCTQTEVIATLKNGREA*CLDPEAPLVQKIVQKMLKGVPK  3. Human GROa(CXCL1) MARAALSAAPSNPRLLRVALLLLLLVAAGRRAAGASVATELRCQCLQTLQGIHPKNI*QSVNVK SPGPHC AQTEVIATLKNGRKA*CLNPASPIVKKIIEKMLNSDKSN 4. Rat MIP-2 (CXCL2)  MAPPTRQLLNAVL*VLLLLLATNHQGTGVVVASELRCQCLTTLPRVDFKNI*QSLTVTPPGPHC AQTEVIATLKDGHEV*CLNPEAPLVQRIVQKILNKGKAN 5. Mouse MIP-2 (CXCL2)  MAPPTCRLLSAAL*VLLLLLATNHQATGAVVASELRCQCLKTLPRVDFKNI*QSLSVTPPGPHC AQTEVIATLKGGQKV*CLDPEAPLVQKIIQKILNKGKAN 6. Human GROB (CXCL2) MARATLSAAPSNPRLLRVALLLLLLVAASRRAAGAPLATELRCQCLQTLQGIHLKNI*QSVKVK SPGPHC AQTEVIATLKNGQKA*CLNPASPMVKKIIEKMLKNGKSN 7. Human IL8 (CXCL8)  MTSKLAVALLAAFLISAALCEGAVLPRSAK ELRCQCIKTYSKPFHPKFIKELRVIESGPHC ANTEIIVKLSDGREL*CLDPKENWVQRVVEKFLKRAENS 8. Human CXCL6  MSLPSSRAARVPGPSGSLCALLALLLLLTPPGPLASAGPVSAVLTELRCTCLRVTLRVNPKTIGK*LQVFPAGPQCSKVEVVASLKNGKQV*CLDPEAPFLKKVIQKILDSGNKKN 9. Mouse IP10 (CXCL10) MNPSAAVIFCLILLGLSGTQGIPLARTVRCNCIHIDDGPVRMRAIGKLEIIPASLSCPRVEIIATMKKNDEQRCLNPESKTIKNLMKAFSQKRSKRAP 10.Human I-TAC  (CXCL11) MSVKGMAIALAVILCATVVQGFPMFKRGRCLCIGPGVKAVKVADIEKASIMYPSNNCDKIEVIITLKENKGQRCLNPKSKQARLIIKKVERKNF 11.Mouse MIP1a  (CCL3) MKVSTTALAVLLCTMTLCNQVFSAPYGADTPTACCFSYSR*KIPRQF*IV*DYF*ETSSLCSQPGVIFLTKRNR*QIC ADSKETWVQEYITDLELNA a b c d Figure 2  Structural homology between PGP and the neutrophil chemokines. ( a ) Alignment of a collagen fragmentisolated from arthritic inflammatory tissue and possessing the PGP sequence 19 with similar sequences contained inneutrophil-attracting ELR + CXC chemokine family members. Chemokines shown are rat, mouse and human CXCL1,CXCL2 and CXCL3 and human CXCL8 and representative ELR – CXCLs and CCLs lacking activity on PMN CXCR1and CXCR2. ELR sequences are green, the GP-containing sequences from collagen and the chemokines are red, andstructural cysteines are blue. The nuclear magnetic resonance imaging solution structure ( b ) of PGP (referring to theN-Acetyl-PGP molecule 22 ) is similar to the SGP motif found in the structure of human IL-8 (ref. 21) and relatedchemokines ( c ). This motif is required for chemokine binding and neutrophil activation 20 , and is shown ( d ; yellow)to be solution accessible in the IL-8 structure and close in space to the important ELR motif (green) present in theneutrophil specific CXC chemokines. For reference, the structural cysteines are shown in red. ARTICLES 318  VOLUME 12  [  NUMBER 3  [  MARCH 2006  NATURE MEDICINE    ©   2   0   0   6   N  a   t  u  r  e   P  u   b   l   i  s   h   i  n  g   G  r  o  u  p    h   t   t  p  :   /   /  w  w  w .  n  a   t  u  r  e .  c  o  m   /  n  a   t  u  r  e  m  e   d   i  c   i  n  e  also completely suppressed neutrophil chemotaxis, whereas eachantibody alone partially suppressed migration ( Fig. 3b ). This effectwas dose dependent, with a 50% inhibitory concentration (IC 50 ) of approximately 0.4  m g/ml ( Fig. 3c ), consistent with IL-8 dosages andmanufacturer’s specifications. We then assayed production of super-oxide resulting from CXCR1 ligation 23 , and showed that both IL-8 andPGP cause release of superoxide from human PMNs, with PGPpretreatment desensitizing cells to IL-8– but not formyl Met-Leu-Phe (fMLP)-dependent production of superoxide, indicating theaction of PGP on CXCR1 ( Supplementary Fig. 1  online).Assaying the presumptive binding of PGP to CXCR1 and CXCR2,we performed receptor competition assays with  125 I-radiolabeledIL-8 on cells from the rat basophilic leukemia (RBL) cell line stably transfected with the CXCR1 and CXCR2 receptors 24 . Excess unlabeledIL-8 markedly suppressed the specific binding of radiolabeled IL-8,and this specific binding is also blocked in the presence of unlabeledPGP ( Supplementary Fig. 1  online). Because the CXCR1- andCXCR2-transfected RBL cells are each chemotactic for IL-8 (ref. 24),PGP can likewise act as a chemoattractant for both, whereas vector-only RBL transfectants are unresponsive to PGP and IL-8 ( Fig. 3d ).The PGG peptide had no effect on migration of CXCR transfectants.Collectively, these data show that, like IL-8, PGP can act on bothCXCR1 and CXCR2.Testing whether PGP’s activity was dependent on chemokinereceptors  in vivo , we intratracheally administered PGP to  Cxcr2 –/– mice 6 (on the BALB/c background). In contrast to normal BALB/cmice, the PGP-treated  Cxcr2 –/– mice did not accumulate PMNs in theairway compared to PBS-treated mice ( Fig. 3e ). PGP is produced  in vivo   in the lung after exposure to LPS We attempted to measure PGP in BAL samples from inflamed airways,obtained at different times after exposure to aerosolized LPS. By developing a highly sensitive electrospray ionization–liquid chroma-tography–mass spectrometry (ESI-LC-MS/MS) protocol, we were ablespecifically detect and quantify the amount of PGP.  Supplementary Figure 2  online shows the ESI-LC-MS/MS profile for a PGP standardand that of the peptide produced endogenously in response to LPS.PGP eluted from liquid chromatography at 6 min with a molecularweight of 312, and was fragmented in a collision chamber to yielddaughter ions with molecular weights of 112 and 140, which can beused to generate a standard curve ( Supplementary Fig. 2  online).Monitoring this profile let us quantify PGP in BAL fluids taken afterexposure to LPS, whereas BAL fluids from unchallenged mice did notcontain PGP ( Supplementary Fig. 2  online).The chronology of cells and mediators involved in the neutrophilicresponse to airway exposure to LPS progressed as follows ( Fig. 4a ). Inthe first 6 h, neutrophil chemokines KC and MIP-2 (CXCL1 andCXCL2, respectively) were produced at high levels, and neutrophilshad begun to infiltrate the airway. Between 6 and 12 h after exposure,PMN cell numbers continued to increase though chemokine concen-trations declined substantially. During this time, the PGP peptidebecame detectable in nanogram amounts, and 12 h after LPS expo-sure, PGP concentrations had risen to more than 800-fold molarexcess to KC. To estimate the  in vivo  PGP concentration in the airway surface liquid, we used a previously described technique 25 , comparingthe concentrations of freely permeant urea in BAL fluid and serum,and found the dilution factor of BAL to be B 77-fold, indicating thatPGP was present at a peak concentration of  B 231 ng/ml or B 0.7  m M(KC is B 7 ng/ml or B 0.8 nM at this time point). The neutrophilnumbers and PGP concentrations remained high until the 24 h timepoint. From 24 to 48 h after exposure, airway neutrophil numbersdeclined as the PGP peptide signal disappeared. PGP peptides in the LPS-treated lung act on PMNs To determine whether the PGP detected in mouse BAL fluids con-tributed to the recruitment of PMNs, we monitored clearance of PGPfrom the airway after intratracheal administration of 100  m g of peptide, and observed rapid depletion of free PGP in BAL fluidswithin 2 h ( Supplementary Fig. 3  online). At a time when PMNinflux was observed (that is, 6 h after PGP treatment;  Fig. 1a ), peptideconcentration was similar to that seen in response to inhalation of LPS (5 versus 3 ng/ml, respectively). To directly test the biologicalactivity of PGP in this model, we evaluated chemotactic activity of BAL fluids from LPS-exposed mice on neutrophils. We used mono-clonal antibodies specific for PGP and chemokines to differen-tially block chemotaxis. The monoclonal antibody 9A4 (refs. 19,26) Media0.1110100IL-810100012345 CXCR1 CXCR2 VectorPGPPGG    C   h  e  m  o   t  a  c   t   i  c   i  n   d  e  x 4 *** ***** 3210IL-8 (50 ng/ml)Anti-CXCR1+2 – – –+ +++ + + + +++++ + – – – – – – – – – – – – – –    C   h  e  m  o   t  a  c   t   i  c   i  n   d  e  x   C   h  e  m  o   t  a  c   t   i  c   i  n   d  e  x PGPCtrl AbAnti-CXCR1Anti-CXCR2100806040200    P  e  r  c  e  n   t  c  o  n   t  r  o   l  c   h  e  m  o   t  a  c   t   i  c  a  c   t   i  v   i   t  y +Anti-CXCR2 ( µ g/ml)PBSPGP Cxcr2  –/– WT876543210    A   i  r  w  a  y   P   M   N  s   (   i  n   d  e  x  e   d   t  o   P   B   S   ) a cedb Figure 3  PMN chemotaxis to PGP is dependent on the CXC chemokine receptors. ( a , b ) PMNs were incubated with 1  m g/mlmonoclonal antibody against CXCR1, CXCR2, both or 2  m g/ml IgG2a isotype control for 1 h at 22  1 C before assayingchemotaxis to 50 ng/ml IL-8 ( a ) or 100  m g/ml PGP ( b ). ( c ) The dose-response curve for the inhibition of PGP-mediatedPMN chemotaxis with monoclonal antibodies against both CXCR1 and CXCR2. ( d ) RBL cells stably transfected to expressCXCR1 or CXCR2 migrate to PGP ( m g/ml as indicated,  P  o 0.05 for CXCR1 and CXCR2 at 10 and 100) and IL-8(50 ng/ml), but not PGG, whereas vector-transfected controls are unresponsive to either ligand. ( e )  Cxcr2  –/– mice andBALB/c control mice ( n   ¼  5 per group) were intratracheally administered 250  m g PGP in 50  m l PBS or PBS alone. BALand cell counts were performed 5 h later; because  Cxcr2  –/– mice are smaller than controls, PMN counts are shown indexedto the values for PBS-treated mice from each group, rather than absolute PMN numbers, which are (left to right) 5.3 and 34.9 for wild-type mice and2.1 and 2.4 for knockout mice (PMN    10 4 per ml BAL fluid). * P  o 0.05, ** P  o 0.01 compared to controls by  t  -tests in  a , d  and  e ; ANOVA in  b . ARTICLES NATURE MEDICINE  VOLUME 12  [  NUMBER 3  [  MARCH 2006  319    ©   2   0   0   6   N  a   t  u  r  e   P  u   b   l   i  s   h   i  n  g   G  r  o  u  p    h   t   t  p  :   /   /  w  w  w .  n  a   t  u  r  e .  c  o  m   /  n  a   t  u  r  e  m  e   d   i  c   i  n  e  recognizes the collagen sequence GPPGPQ, which is only exposedafter proteolysis of collagen, and specifically inhibits PGP chemotacticactivity. Coincubation of 9A4 with chemoattractants potently blockedchemotaxis to PGP, but not to IL-8, KC or MIP2 ( Supplementary Fig. 3  online).A 1-h preincubation of BAL fluids taken 24 h after exposure to LPSwith either 9A4 or monoclonal antibody specific for KC and MIP-2markedly reduced ( B 40%) chemotaxis to the BAL fluid, and whenthe antibodies were used in combination, chemotaxis was furthersuppressed ( B 65% decrease;  Fig. 4b ), suggesting that PGP-containingpeptides and chemokines independently and additively affect PMNtraffic 24 h after LPS challenge.We then tested whether 9A4 or monoclonal antibody specificfor MIP-2 and KC could block LPS-dependent neutrophil infiltrationof the lung  in vivo . These antibodies, when intratracheally adminis-tered 1 and 10 h after exposure to LPS, similarly reduced theneutrophil burden in the lungs of mice 24 h after exposure to LPScompared to appropriate isotype control antibody or PBS-treatedmice ( Fig. 4c ). Chronic airway exposure to PGP causes inflammatory remodeling Chronic airway inflammatory diseases like COPD manifest alveolarenlargement and right ventricular hypertrophy. To evaluate whetherthe presence and activity of PGP in inflammatory situations caninitiate a cascade of events culminating in pathology, we chronically administered PGPor PBS to the airways of mice and evaluated clinicalparameters of lung pathology.We treated C57 BL/6 mice with 250  m g PGP by direct airway instillation twice weekly for 12 weeks. We evaluated lungs and hearts 2weeks after the last exposure for pathological changes.  Figure 5  showsmicrographs of representative lung sections taken from PBS-treatedcontrol ( Fig. 5a ) and PGP-treated ( Fig. 5b ) mice. We observedenlargement of alveoli, as the mean alveolar linear intercept (Lm)was significantly increased (by 21%,  P  o 0.05) in PGP-exposed mice( Fig. 5c ); these changes are similar to those reported for C57 BL/6mice exposed to smoke from two cigarettes daily 6 d per week for6 months 27 . BALB/c mice showed statistically significant butless substantial increases in Lm after 8 weeks of PGP treatment( Supplementary Fig. 3  online). BALB/c and C57 BL/6 mice alsoshowed increased levels of vascular endothelial growth factor (VEGF)in the lung (data not shown) and right ventricular hypertrophy ( Fig. 5d  and  Supplementary Fig. 3  online) in response to PGPtreatment; at present, we do not know whether this effect is aconsequence of PGP’s recruitment of PMNs or a secondary effect of PGP. It remains to be tested whether specific inhibition of PGP by monoclonal antibody 9A4 would prevent the inflammatory andstructural changes seen in response to long-term exposure to peptide. Figure 4  Kinetics of neutrophil influx, CXCchemokine and PGP production in the mouseairway after exposure to aerosolized LPS ( a ).Monoclonal antibodies against KC and MIP-2 orPGP inhibit i n vitro   PMN chemotaxis caused bybronchoalveolar lavage fluids from LPS-exposedmice ( b ), and decrease the  in vivo   neutrophilburden in lungs of LPS-exposed BALB/c mice( c ). ( a ) Mice were exposed to 100  m g/mlaerosolized LPS for 1 h at time 0. BAL samples( n   ¼  3 per time point) were collected andanalyzed for neutrophil number (green squares,outer left axis; numbers are percentage of PMNsof total cells); KC and MIP-2 chemokineconcentrations by ELISA (blue diamonds andblack triangles, respectively, inner left axis); andPGP (red circles, right axis) by ESI-LC-MS/MS.( b ) Twenty-four hours after LPS exposure, BALsamples were collected, and supernatants wereincubated for 1 h with monoclonal antibodiesagainst KC (50  m g/ml), MIP-2 (50  m g/ml), PGP(9A4, 5  m g/ml) or isotype control antibody (50  m g/ml) before PMN chemotaxis. Scale shows activity of BAL samples from LPS-exposed mice versus LPS-naive mice, showing mean ± s.e.m. from six mice. ( c ) Mice exposed to aerosolized LPS were intratracheally administered PBS, isotype control monoclonalantibody (50  m g total), monoclonal antibody against MIP-2 and KC (25  m g each) or monoclonal antibody 9A4 (30  m g) in 35  m l PBS 1 and 10 h after LPS.PMNs were counted from 24 h BAL samples ( n   ¼  3 per group); isotype controls IgG2a (third bar) and IgG1 (fourth bar) are controls for MIP2- and KC-specific antibody and 9A4 antibody, respectively. * P  o 0.05, **  P  o 0.01 compared to positive controls by ANOVA for  b  and  c .    B   A   L   f   l  u   i   d   P   M   N  s   (  c  e   l   l  s   /  m   l    ×     1   0    4    ) 050100150200250 06 h12 h24 h48 h    B   A   L   f   l  u   i   d   C   X   C   L  c  o  n   t  e  n   t   (  p  g   /  m   l   ) (93.4%)(84.9%)(78.1%)(53.1%) Aerosol LPS exposure0123    B   A   L   f   l  u   i   d   P   G   P   (  n  g   /  m   l   ) KCMIP-2PMNsPGP   (1.2%) 050100150200250    P  e  r  c  e  n   t  c   h  e  m  o   t  a  x   i  s  a   b  o  v  e  n  o  r  m  a   l 10080    A   i  r  w  a  y   P   M   N  s   /  m   l    ×     1   0    4 6040200Control BAL fluid+––––––+++++––+––––––+–+––––++24 h BAL fluidControl IgG mAbControl IgG mAbLPS exposuremAb 9A4––––+–––+–––+–––++––++++mAb 9A4Anti-MIP2/KC mAbAnti-MIP2/KC mAb a bc 50403020100 ****** 806040    L  m    (      µ   m   )   R   V   /   (   L   V  +   S   ) 200PBS PGP PBS PGP0.30.20.10 * * a bc d Figure 5  Extended exposure to PGP causes alveolar enlargement and rightventricular hypertrophy. C57Bl/6 mice were intranasally treated with PGP(250  m g in PBS) or PBS twice weekly for 12 weeks, and evaluated 2 weeksafter the last treatment. H&E-stained lung sections from PBS- ( a ) andPGP-treated mice ( b ) were evaluated for alveolar enlargement. Scale bar,100  m m. ( c ) A 21% increase in Lm in the PGP-treated group (* P   ¼  0.021).( d ) Right ventricular (RV) mass is proportionally greater than the rest of thelower heart (that is, the left ventricle (LV) and the septum (S)) for PGP-treated mice than for PBS-treated controls (22%, * P   ¼  0.025). ARTICLES 320  VOLUME 12  [  NUMBER 3  [  MARCH 2006  NATURE MEDICINE    ©   2   0   0   6   N  a   t  u  r  e   P  u   b   l   i  s   h   i  n  g   G  r  o  u  p    h   t   t  p  :   /   /  w  w  w .  n  a   t  u  r  e .  c  o  m   /  n  a   t  u  r  e  m  e   d   i  c   i  n  e  PGP is present in BAL fluids from people with COPD To determine whether PGP is a marker of pathology in human disease,we screened for PGP in BAL fluid samples from individuals withCOPD. Using the same ESI-LC-MS/MS method with more sensitiveinstrumentation, we could quantitatively detect as little as 10 pg/mlPGP ( Fig. 6 ). Three of the five samples from individuals with COPDtested were positive for PGP, with concentrations averaging 363 pg/ml,whereas the 2 samples from the 18 control samples with signal abovethe detection limit averaged only 22 pg/ml. The difference in PGPconcentration between positive samples of the groups was highly significant ( P   o  0.01 by unpaired  t  -test with Welch correction),whereas the incidence of PGP detection was also significantly differentbetween the groups ( P  ¼ 0.014 by Fisher exact test). Other importantfeatures of this study were that the three individuals with COPD whowere positive for PGPalso had evidence of emphysema by CTscan andlower mean FEV1 scores (50 ± 6% versus 62% predicted) than the twoindividuals with COPD who were negative for the peptide. DISCUSSION Neutrophil chemoattractant properties of PGP-containing peptideshave been documented since 1970 (refs. 7,28). We report here for thefirst time that this activity occurs through the neutrophil CXCchemokine receptors. We also show that tissue breakdown afterexposure to LPS causes generation of PGP  in vivo . Furthermore,collagen-derived peptides in BAL fluids are active PMN chemoattrac-tants, and specifically targeting these with monoclonal antibody 9A4relieves the neutrophil burden in the lung 24 h after exposure to LPSto about the same extent as treatment with chemokine-specificmonoclonal antibody. PGP exposure studies show that the novelactivity of this peptide can recruit PMNs into airways and initiateinflammation, causing alveolar enlargement. Finally, the presence of PGP in samples from individuals with COPD suggests that collagenprocessing and the resulting recruitment of PMNs may actively drivethe pathology of this disease.Several ECM proteins contain the PGP motif, which is mostabundant in collagen but is also present in collagen-like domains of elastin and surfactant proteins A and D. Because collagen normally forms a helical trimer, cleavage by collagenase (MMP-1 or MMP-9)exposes the normally hidden PGP-containing strands 19 , unmaskingproinflammatory activity of structural proteins. Generation of thetripeptide described here may require further cleavages by collagenasesor gelatinases, and modification by acetylases present in inflammation.Because PGP is detected after early PMN infiltration, it is unclearwhether neutrophil-derived enzymes are required for production of PGP, and the specific sequence of molecular events creating thischemoattractant remain unknown.Macrophages release several MMP species after inflammatory stimuli, and their presence and activity correlate with COPD pathol-ogy  29 . Models of exposure to cigarette smoke show that both break-down of ECM (by neutrophil or macrophage elastase) andaccumulation of neutrophils are required for pathological change 30 ,and mice lacking either macrophage elastase 31 or neutrophil elastase 27 do not break down ECM and do not recruit PMNs into the lung,despite an intact chemokine signaling system. Our data suggest thatPGP and related peptides are active in acute inflammation and may recruit cells in chronic situations where MMPs are upregulated andcells continue to infiltrate despite a lack of chemokine signal.The basis for PGP’s activity lies in its molecular similarity to the GPmotif present in all ELR  + CXC chemokines. Mutation or juxtaposition(GP to PG) of these residues in IL-8 render the molecule essentially nonfunctional 20 . This motif was suggested to be the probable receptorbinding site in primary IL-8 studies 21 , whereas other reports suggestthat this motif initially binds the receptor, allowing cooperativebinding between other parts of the ligand and receptor to efficiently transduce G-protein signaling 32 . This cooperativity probably accountsfor the pharmacological differences between PGP and the chemokines,whose  in vitro  activities are optimal at micromolar and nanomolarconcentrations, respectively. As seen with other peptides and cyto-kines 33–35 , doses of PGP or chemokines required to overcome lungclearance mechanisms and observe airway inflammatory changes arenearly 1,000-fold higher than amounts in BAL fluids, whereas the doseratio between PGP and chemokines is consistent between  in vivo  and in vitro  studies. We thus consider PGP a partial agonist of chemokinereceptors 36 , which could be present and perhaps active in multipleinflammatory settings where matrix breakdown is occurring.Antagonizing chemokines and leukotrienes incompletely block neutrophil chemotaxis to human COPD samples 37 . Detection of PGP in samples from individuals with COPD could explain residualchemotactic activity, and may account for long-term sequelae seen inchronic inflammation. Although the concentration of PGP in the BALfluid is below its chemotactic threshold, it becomes more relevantwhen we account for the B 1:100 dilution of airway lining fluid by thelavage process reported in earlier human studies 25 , placing the  in situ concentration of diffusible PGP around 0.1  m M. Larger peptidescontaining the PGP sequence possess similar chemotactic activity  9,38 and could also contribute to overall chemotaxis, though such specieshave not been measured. Future longitudinal studies in humans formultiple PGP-related peptides, perhaps with breath-condensatesamples 39 , will probably best determine the contribution of thispathway to airway disease and use as a disease marker.Clinical findings with IL-8–specific monoclonal antibodies 40 inindividuals with COPD implies that cell trafficking is a realistictherapeutic target for inflammatory diseases. The activity of PGP asa PMN chemoattractant may represent another target. This idea isconsistent with the protease-antiprotease hypothesis of airway diseases 41 , and is corroborated by assertions that breakdown of ECM may be an important therapeutic target in chronic airway diseases 29 , and that tissue damage is self-sustaining 42 . PGP signalingthrough CXCRs on neutrophils indicates that combining MMPblockade with chemokine receptor antagonism may dampen multiple 0100200300400    B   A   L   f   l  u   i   d   P   G   P   (  p  g   /  m   l   ) COPD subjectsControl subjectsDetection limit P   = 0.014 Figure 6  BAL fluid samples from individuals with COPD contain elevatedlevels of PGP. BAL samples from humans were assayed for PGP usingESI-LC-MS/MS. Incidence of PGP is greater among individuals with COPD(3 of 5) than controls (2 of 18,  P   ¼  0.014), and detected levels of PGPare higher in individuals with COPD (363) than controls (22;  P   ¼  0.015). ARTICLES NATURE MEDICINE  VOLUME 12  [  NUMBER 3  [  MARCH 2006  321    ©   2   0   0   6   N  a   t  u  r  e   P  u   b   l   i  s   h   i  n  g   G  r  o  u  p    h   t   t  p  :   /   /  w  w  w .  n  a   t  u  r  e .  c  o  m   /  n  a   t  u  r  e  m  e   d   i  c   i  n  e
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks